Open iTunesconnect, go to Analytics for your App > Metrics > Number of Sessions App > Metrics > Crashes Divide Crashes into Sessions to get a percentage.
If your game is played 100 sessions and has three crashes, then your app crashes 3% of the time (based on opt in analytical data from iTunes Connect). If your game is played 1,000,000 sessions and crashes 30,000 times, your app crashes 3% of the time. 3% of the time is 1 out of 33 chance of crash 5% of the time is 1 out of 20 7% of the time is 1 out of 14
This is not necessarily true. A crash can be an error which is recorded but doesn't cause the app to crash and go to the home screen. In any application there are crashes all the time but most of these are caught by error handlers written in the code. You can see this happen by running your app on a device connected to Xcode or Eclipse. You may see numerous errors happening in logcat for instance but the game doesn't crash on the device.
You can check the local crashes also inside the device (=> Privacy ->Diagnostics & Usage -> Diagnostics & Usage Data). Interestingly, the iTunesConnect data is "feedback" from the customers, so checking local data of one or a few devices and comparing those to iTC data might lead to a large deviation
No, for iOS, a crash is the app quitting. This is a helpful link from Apple: https://developer.apple.com/library...lyzingCrashReports/AnalyzingCrashReports.html
I stand corrected however, this is a very interesting topic that I will investigate a bit more. I find it hard to believe my app crashed thousands of times without a) it ever happening to me b) someone complaining about it.
unfortunately, there are so many games out there that if your app crashes people will move on to play another game, (and also not share your game to others) If you can open Xcode > Window > Organizer you can see crashes from the App Store version...here is a screenshot of my game in which you can see 70% of my crashes are related to OpenAudio and I've witnessed crashes on random defeated animations which makes me think the code to load and unload audio assets has a bug (ticket submitted and fingers crossed they can fix it)